Apparent Asymmetry in Fingerprint Similarity Searching is a Direct Consequence of Differences in Bit Densities and Molecular Size

Wang, Yuan; Eckert, Hanna; Bajorath, Jürgen

doi:10.1002/cmdc.200700050

Cited by 26 publications

(41 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Furthermore, Tc was shown to be the coefficient of choice in similarity searching when reference and database molecules had comparable complexity, but when these complexity levels varied, other coefficients were preferred [76]. Finally, an apparent asymmetry of Tversky similarity values observed in database searching [77] could also be directly attributed to complexity effects [78]. Clearly, circumventing complexity effects is not only of scientific interest but also of high relevance for practical applications.…”

Section: Circumventing Intrinsic Limitations: Complexity Effectsmentioning

confidence: 95%

“…As a complexity-independent fingerprint representation, PDR-FP has been reported [32] that generates a constant bit density of 93/500 for all compounds, as described above. Thus, similarity search calculations with PDR-FP are not biased by differences in molecular complexity [78] and have been shown to be particularly effective on compound classes having high structural diversity where other types of fingerprints produce only low compound recall [32,80]. Furthermore, it has recently also been demonstrated that conventional keyed fingerprints can be rendered complexity-independent through 'balanced codes' transformation, that is, by merging a fingerprint with the complement of its bit setting, which generates a constant bit density of 50% (but doubles the size of the fingerprint) [81].…”

Section: Circumventing Intrinsic Limitations: Complexity Effectsmentioning

confidence: 99%

See 1 more Smart Citation

Advances in 2D fingerprint similarity searching

Geppert¹,

Bajorath²

2010

Expert Opinion on Drug Discovery

Self Cite

View full text Add to dashboard Cite

The reader is provided with an overview of the state-of-the-art fingerprint design and search strategies developed. It will be possible to rationalize opportunities and limitations of 2D fingerprint similarity searching. Take home messages: Fingerprint search calculations are more complex than it might appear at first glance and susceptible to complications that are often overlooked in practical applications. Fingerprint search performance typically only depends on relatively small subsets of bit positions. Recently, different fingerprint engineering strategies have been applied to 'tune' existing fingerprints in a compound class-directed manner. Fingerprints have substantial scaffold hopping potential, despite the simplicity of their design.

show abstract

Section: Circumventing Intrinsic Limitations: Complexity Effectsmentioning

confidence: 95%

Section: Circumventing Intrinsic Limitations: Complexity Effectsmentioning

confidence: 99%

Advances in 2D fingerprint similarity searching

Geppert¹,

Bajorath²

2010

Expert Opinion on Drug Discovery

Self Cite

View full text Add to dashboard Cite

show abstract

“…In addition to uncertainties associated with the interpretation of calculated similarity values, molecular complexity or size effects are known to bias fingerprintbased similarity evaluation and negatively affect search performance [6,22,42,43]. In a milestone publication, Flower demonstrated that reference molecules of increasing size generate systematically higher Tc values in databases searching [6].…”

Section: Intrinsic Caveatsmentioning

confidence: 97%

Advanced Fingerprint Methods for Similarity Searching: Balancing Molecular Complexity Effects

Wang¹,

Bajorath²

2010

CCHTS

Self Cite

View full text Add to dashboard Cite

Fingerprint representations of molecular structure and properties were among the first computational tools for similarity searching and are widely applied to this date, which is in part due to their computational efficiency and ease-of-use. Moreover, despite their simplicity, 2D molecular fingerprints have been surprisingly successful in identifying novel active compounds. However, properly applying 2D fingerprint representations and similarity metrics in the search for active molecules is much more complicated than one might assume. For example, no generally applicable similarity functions and threshold values exist that reliably indicate a level of molecular similarity that results in similar biological activity. Furthermore, fingerprint search calculations are known to be biased by molecular size and complexity effects, which often lead to a significant increase in false-positive rates. In this contribution, we will describe known caveats associated with fingerprint similarity searching and strategies to overcome these difficulties including the design of complexity-independent fingerprint representations and similarity metrics.

show abstract

“…The Tversky coefficient can be biased towards substructure similarity. The Tversky coefficient is defined as virtual screening recall -a value which has been found to also be beneficial in similarity searching using fingerprints 28,29 , and has also shown potential in scaffold hopping with fingerprints 30 . …”

Section: Repeat Steps 2 and 3 Until The Set Is Emptymentioning

confidence: 99%

Maximum Common Substructure-Based Data Fusion in Similarity Searching

Duesbury

Holliday

Willett

2015

J. Chem. Inf. Model.

View full text Add to dashboard Cite

Data fusion has been shown to work very well when applied to fingerprint-based similarity searching, yet little is known of its application to Maximum Common Substructure (MCS)-based similarity searching.Two similarity search applications of the MCS will be focussed on here. Typically, the number of bonds in the MCS, as well as the bonds in the two molecules being compared, are used in a similarity coefficient. The power of this technique can be extended using data fusion, where the MCS similarities of a set of reference molecules against one database molecule are fused. This "group fusion" technique forms the first application of the MCS in this work. The other application is that of the chemical hyperstructure. The hyperstructure concept is an alternative form of data fusion, being a hypothetical molecule that is constructed from the overlap of a set of existing molecules. This paper compares fingerprint group fusion (extended-connectivity fingerprints), MCS similarity group fusion, and hyperstructure similarity searching, and describes their relative merits and complementarity in virtual screening. It is concluded that * To whom correspondence should be addressed 1 the hyperstructure approach as implemented here is less generally effective than conventional fingerprint approaches.

show abstract

Apparent Asymmetry in Fingerprint Similarity Searching is a Direct Consequence of Differences in Bit Densities and Molecular Size

Cited by 26 publications

References 13 publications

Advances in 2D fingerprint similarity searching

Advances in 2D fingerprint similarity searching

Advanced Fingerprint Methods for Similarity Searching: Balancing Molecular Complexity Effects

Maximum Common Substructure-Based Data Fusion in Similarity Searching

Contact Info

Product

Resources

About